Miles Brundage: shipping with known flaws isn't iterative deployment, it's just faster shipping
Miles_Brundage · x · 2026-10-03
Former OpenAI policy head Miles Brundage weighed in on the limits of "iterative deployment." He concedes that some lessons can only be learned by shipping to real users. His criticism targets companies that ship with already-known flaws — a practice that doesn't actually accelerate learning, only the pace of shipping. The post is widely read as a veiled critique of current frontier-lab release cadence.
Related event: Ex-OpenAI's Brundage Criticizes Industry Safety as Failing(4 posts)→
More from AGI Musings
- Anthropic researcher: talk of Claude having a soul may seed machine consciousness — ibab · 2026-10-04
- 80% who say AI will stay a chat interface truly believe it, 20% self-soothe — danfaggella · 2026-10-04
- Researcher: as AI outperforms humans, alliances with biological intelligence matter more — begusgasper · 2026-10-04
- Alignment training may backfire: policing chain-of-thought could make AI intelligence jagged — robleclerc · 2026-10-04
- AI evolved from Wright brothers to mass aviation in years, safety lessons won't have time to land — soumitrashukla9 · 2026-10-04
- Bindu Reddy: stop testing AI like humans with agency — they're tools we can unplug — bindureddy · 2026-10-04